Anatomy of an Integration Recovery
Contracts uninvoiced for six months, taxes unpaid, drivers stopped by the police. Inside a GE integration recovery, and the method that saved the business.
There is a moment in many integrations when leadership quietly acknowledges that it is not working. Sometimes the signals are slow: a synergy tracker nobody trusts, workstreams on hold, two legacy organizations settled into cold coexistence. And sometimes the signals are brutal. I lived the brutal version at GE, and it taught me more about integration recovery than any playbook.
A Merger of Two Verticals
GE merged two businesses from different verticals: one doing fleet management, the other car financing. A new CEO and leadership team were named, and the acquired business began its migration onto the new platform. On paper, a sensible consolidation.
Two issues plagued the migration. First, an error in the lease calculation made every invoice wrong, and the clients, mostly multinational corporations with strict payment controls, could not pay us. Second, the accounts payable module of the ERP had not been implemented properly, so the business could not pay many of its own invoices, including vehicle taxes and payments to car dealers.
Either issue alone was serious. Together they meant the business could neither collect money nor pay it out. And then came the aggravating factor that turns problems into crises: for a few months, the issues were hidden. The local CEO and the European IT leader each pointed at the other, playing ping pong with responsibility while the backlog grew. By the time the problem surfaced to the European CEO, some contracts had gone uninvoiced for six months.
The consequences had already reached the street. Customer drivers were being stopped by the police because the tax on their GE vehicle had not been paid. Car dealers stopped delivering cars because they had not been paid. The European CEO believed we would lose the business entirely, and a year later the episode cost her her job. That is what a failed migration looks like when it is left to fester: not a red status on a slide, but drivers explaining themselves to the police and suppliers cutting you off.
Two Days to Diagnose, One Week to a Plan
The day the issue surfaced, I was dispatched to Sweden. Not a committee, not a request for a report by month end. Same day, on a plane.
The first visit was a two-day diagnostic, reported immediately. I came back the following week with the essentials: the list of top issues, their priority and dependencies, an action plan to fix them, and the resources I needed. Exactly one week after arriving in a business I had never seen, the CEO and the key stakeholders had agreed the recovery plan, and I was tasked to lead it with full authority over the leadership team. My priorities superseded all other business priorities, explicitly.
Note what did not happen. There was no four-week study, no perfect analysis, no hundred-page deck. The plan was directionally right and knowingly incomplete, built to be fine-tuned in execution. In a crisis, a good plan agreed in a week beats a perfect plan agreed in a quarter, because the damage compounds daily and so does the organization’s despair.
Boots on the Ground
My first request was unusual: relocate immediately a full IT team from the UK and India to Sweden. The week after, a small SWAT team arrived: subject matter experts, lean six sigma, IT.
The efficiency argument for colocation is real, but the leadership message mattered more. We come here full time to solve the issues with you. Not remotely, not by video call, not through tickets and escalations. Having everyone around the same table, sharing the same urgency, proved to be the breakthrough. Problems that had survived six months of remote finger-pointing did not survive days or weeks of people sitting together.
The proof came quickly. The main issue with the rent calculation turned out to be almost trivial. Identified and fixed in short order, once the right people sat in the same room with a single mandate, after six months in which it had defeated the organization. That is worth pausing on: the technical problem was never the hard part. The operating model around it, split responsibilities, remote teams, hidden bad news, was what kept it alive.
From Survival to Momentum
With the calculation fixed, all invoices were reissued. An action plan secured the top clients one by one, and in the end only one left for the competition, in a business the European CEO had expected to lose entirely.
Then the recovery turned into an upgrade. We moved the ERP front end to the platform used by all other GE entities in Europe, one I was able to accelerate because I had led its original implementation. Customers got a modern self-serve platform, and their satisfaction rose above where it had been before the crisis. Later, process experts from several European entities spent days each week transferring knowledge and best practices to the local teams.
That last phase did something no communication plan could have done. The teams felt something was happening and that they were part of it. Employee attrition went down, in the middle of what had been a death spiral. The business was saved, and it came out stronger than it went in.
The Lessons: A Recovery Method
Looking back, the recovery worked because of a method, and the method transfers to any stalled integration.
1. Diagnose short
Two days on site tell you more than four weeks of remote analysis. The goal of the diagnostic is not completeness, it is to identify the vital few issues, their dependencies, and the resources needed. Speed of diagnosis is itself a message: this is now the priority.
2. Plan fast, refine in motion
Get a directionally correct plan agreed within days, with explicit priorities and owners, and accept that it will be fine-tuned weekly. Waiting for certainty is how the previous six months were lost.
3. Demand a full mandate
I could move fast because my priorities explicitly superseded all other business priorities, and the organization knew it. A recovery leader with an ambiguous mandate becomes one more voice in the stall. Which decisions the leader makes alone, which go to the committee, and what done looks like must be settled on day one.
4. Move the teams on site
Colocation is a force multiplier and a leadership statement. Remote teams exchange blame; colocated teams exchange solutions. If the people who must fix the problem are not sharing a table and a sense of urgency, the recovery has not actually started.
5. Fix sequentially, biggest pain first
Attacking everything at once achieves nothing. Fixing the invoice calculation first, the issue behind both the cash collapse and the customer crisis, produced visible results within weeks, and those results bought traction and engagement for everything that followed.
6. Install a hard cadence
Progress reported every week to the committee, roadblocks discussed and cleared in the room, clear actions and accountability out of every session. The cadence is what keeps the plan honest and the momentum alive when the initial adrenaline fades.
7. Bring outside expertise, transfer it before leaving
The SWAT team and the visiting process experts brought capability the local team did not have, and then deliberately handed it over. Outside experts who fix and leave create dependency. Outside experts who fix and teach create capability, and the local team’s pride in the recovery is what makes it durable.
8. Leverage momentum
Once the crisis was stabilized, we did not stop, we used the credibility to modernize the front end and upgrade the customer experience. A recovery that merely returns to the pre-crisis state wastes its own momentum. The organization is mobilized, decisions are moving, resistance is low: that is precisely the moment to make the improvements that were impossible before.
The Uncomfortable Lesson About Bad News
One lesson sits above the method. The six months of damage were not caused by the lease calculation error. They were caused by two leaders hiding it while arguing about whose fault it was. The technical fix took weeks; the concealment nearly killed the business.
A strong execution culture does not mean everything goes right. It means issues surface early enough to be fixed. Bad news is not the problem. Late bad news is the problem. Any integration governance that does not actively reward the early surfacing of problems, and visibly punish their concealment rather than their existence, is incubating its own Sweden.
Be the Solution, Not the Problem
Early in the recovery, the European CEO asked me directly: who is responsible for this? I gave a candid answer. IT had most likely mishandled the migration, and the business was paying the price. But the local CEO had lived with the problem for months, never raised it, and never asked for help. Both things were true, and I owed the CEO both.
What happened next said more than any answer I could have given. The European IT director threw himself into the recovery: in Sweden with his team, at the table, staying late into the night through the critical sessions. Whatever had gone wrong on his watch, he was visibly part of fixing it. The local CEO did the opposite: distant, disengaged, at times an actual roadblock for teams that needed decisions. Two weeks later, she was gone.
The lesson has stayed with me ever since. In a crisis, who caused the problem matters far less than who is solving it. Organizations do not judge leaders on the origin of a mess, origins are usually collective and partly bad luck anyway. They judge leaders on their behavior once the mess is visible: present or absent, clearing obstacles or becoming one, owning the fix or protecting their position. The IT director had a share of the responsibility and kept his career, because he chose to be part of the solution. The local CEO may have had less direct fault, and lost hers, because she chose to stand apart from it.
If you find yourself anywhere near a failing integration, that is the choice that defines you. Not whether the problem happened on your watch, but whether the fix did.
The Deal Made the Headlines. The Recovery Made the Returns.
Not every stalled integration should be recovered; sometimes the fact base shows the original deal logic is gone, and the honest move is to re-size the ambition. But in most cases, the value is still there. It was never destroyed, it was simply left undelivered, sitting in unmade decisions, unfixed processes, and hidden problems.
Recovering it is unglamorous work: diagnose in days, decide in a week, put people around a table, fix the biggest pain first, report every week, and keep moving. I have seen a business go from drivers being stopped by the police to higher customer satisfaction than before the crisis, in months, with that method. The plan does not need to be perfect. It needs to be fast, sequenced, resourced, and led.
